How DBZDrums Uses AI

As I mentioned in The Dev Behind DBZDrums, AI has been essential for building this product in the time I actually have available: mostly nights and weekends.

Without that support, doing everything alone would take many, many months before the project could become something useful to people beyond myself.

And when I say “the project”, I do not mean code only. DBZDrums includes the app, the landing page, this blog, prototype testing, idea validation, documentation, releases, languages, assets, and many other small pieces that all need to work together.

This post is a look at how I use AI across some of those areas.

In the DBZDrums app

In the app development process, I keep full control over what should be built and how the main decisions are made.

I use AI agents to speed up parts of the work: writing pieces of code, generating notes about tests that were run, documenting decisions and conclusions, maintaining the roadmap, organizing what is done and what still needs to be done, creating tests, and helping with release management.

I also use AI to turn visual ideas into code. Even though I know the technologies involved, building graphical interfaces has always been harder for me than other parts of software development. Many times I start with hand-drawn sketches or text descriptions, and AI helps me turn those ideas into a first working version.

That does not mean I accept everything automatically.

I ask for incremental improvements and review the work carefully: the generated code, the organization, the documentation, and the impact of the change before I accept anything as part of the project.

I also use AI for critical review, not only for production. I ask it to point out risks, challenge a solution, look for inconsistencies, and raise details I may have missed. That use is valuable because it helps test decisions before they become part of the product.

No code enters the project without a build, test, or human validation step that matches the size and risk of the change.

AI also helps me move faster when researching technical topics, looking for good coding practices, checking whether an approach is viable, answering focused technical questions, and supporting multiple languages.

On the website and blog

The DBZDrums website is built with an open source static site framework.

In that context, AI helps me organize files, format content, correct text I wrote, and prepare versions in other languages. Yes, I write almost all of the text myself. Sometimes I accept small adjustments suggested by the agent, but the idea, direction, and main content still come from me.

I also use AI to help maintain this blog: structuring posts, improving the flow of the text, preparing metadata, and adapting posts into other languages so the project can reach more people.

In release management

Together with AI agents, I wrote tools and instructions to automate a large part of the release process.

That includes creating version control tags, running builds, generating deployment images, and updating the environments where the project components are running.

This kind of automation matters because it reduces repetitive work and lowers the chance of mistakes in steps that need to be handled carefully.

How I do not use AI

It is also important to be clear about how I do not use AI.

I do not use AI to “invent” solutions or features just to make the project look bigger than it is. All product decisions and features are defined by me, based on the problem I am trying to solve and the direction I believe makes sense for DBZDrums.

I also do not use AI as an unrestricted replacement for other people’s creative work.

Whenever possible, I try to value the work of people who create components, libraries, samples, and reusable material. Fortunately, so far that has been very possible.

For the app and website assets, the small amount of AI-assisted work started from sketches and directions defined by me. AI mainly helped me draw or visually refine ideas I would not have been able to produce well on my own, such as the logo and icons.

The samples used for rhythm playback are public-domain, free-to-use samples made available by DrumThrash: Free Drum Samples.

AI does not replace direction

For me, AI is a support tool.

It helps accelerate, organize, review, research, and prototype. But the product direction, final decisions, validation, and responsibility remain human.

DBZDrums only makes sense because it comes from a real problem I experienced while studying drums. AI helps make the project viable in the time I have available, but it does not decide what DBZDrums should become.

That combination of technical experience, a practical problem, and modern tools is what has allowed the project to move forward.

DBZDrums Dev